I am reaching out to you to gain insights on what approach do you guys follow when it comes to Job recording and what do you think are the pros/cons of each approach mentioned below.
Approach 1:
Have the job recording enabled for all processes at all times. So, whenever an issue happens, you can go access the recording to get a view on issue that happened during process run.
Approach 2:
Have it disabled at all times. Whenever a process starts throwing errors, you enable the recording for that process and re-run it to capture the issue.
Kindly feel free to provide a 3rd or more approaches you follow currently or are possible here.
Your inputs will help us greatly to decide how we proceed with using the job recording feature.
Due to recoding feature limitation of recording only last 3 minutes, we use record failed transaction settings.
I would suggest to not use second approach as there could be chance that transaction faulted in initial run might not reproduce the issue to record. So have it recorded at first place.
I agree to have it enabled at all times as there can be instances where we may fail to reproduce that error to record. And having it enabled only to store failed txns would allow to optimally use storage.
Could you also share how you guys are approaching usage of live streaming in your organizations?
Per my understanding of this feature at the moment, I think below:
Live streaming is best suited right after prod implementation as it allows us to also remote control the process if needed and would avoid the need of gaining prod machine credentials to login and debug.
Job recording is something that can be enabled after disabling live streaming and better approach would be to record all failed queue transactions instead of failed jobs as some failed transactions doesn’t lead to failed jobs thus would prevent recording capture to get triggered.
Is there any scenario that would demand both live streaming and job recording to be enabled?
Job recordings are not available after 14 days(basis on how it was enabled), but for live streaming, I am not sure for how long it remains accessible once process is completed. Or is it only available till process gets completed and after that its only available if job recording was also enabled for same process?
We have enabled this for all the bots as whenever we need to monitor any job execution, this comes very handy to see what is bot doing.
Your understanding point 1 is right but in point 2 I would like to correct like this:
We can keep Live streaming and recording on right from the first time coz live stream don’t record anything. It just streams the execution of the bot run when it’s running but when our bot fails we will require recording to identify the bug and fix.
The scenario could be any hypercare bot which just started the production. Here we proactively monitor the runs to see if it’s working fine or not. So to see in action we will be using Live streaming and if something goes wrong we will be using recording.
Job live streaming won’t be available after job run finished.